Forecasts

Forecasts give you a 5-day outlook of expected energy production for every plant you can access, so you can plan operations, anticipate revenue, and spot a plant that is about to underperform before it costs you money.

Forecast Concept

A forecast combines each plant's recent measured behavior with the local weather forecast to estimate how much energy it will generate over the next five days. The estimate is grounded in physics — irradiance, temperature, and a clear-sky model of the site — rather than statistical guesswork. There are no opaque machine-learning predictions: every figure traces back to weather inputs and the plant's own observed performance.

Two complementary views are available:

• Portfolio overview — a card grid showing the 5-day energy range and weather for each accessible plant at a glance.
• Per-plant detail — a chart for a single plant with daily expected energy, a clear-sky reference, and hour-by-hour power for each of the five days.

What the Forecast Delivers

Portfolio Overview

For each plant you have access to, the overview provides a daily forecast for the next five days:

• Expected energy — a mean value with a high and low band (in Wh), so you see the likely range, not just a single number.
• Standard deviation — a one-sigma spread that quantifies how confident the estimate is for that plant.
• Performance ratio — the expected ratio of actual to theoretical production, carried into the forecast from the plant's measured behavior.
• Daily weather — roughly ten weather fields per day (including precipitation, snow, wind, temperature, and a weather code), so you can see why a given day's output is expected to be high or low.

By default the estimate is built from the plant's behavior over the last 45 days; you can widen or narrow that sample window.

Per-Plant Detail

Open a single plant to see a richer five-day picture:

• Daily expected energy — the production estimate for each of the next five days.
• Clear-sky reference — the theoretical production the plant would reach under perfect clear-sky conditions, plotted alongside the forecast so you can read the gap between ideal and expected.
• Hourly power curve — power samples through each day, fixed-scaled against the plant's peak alternating-current capacity so the chart shape is directly comparable day to day.

If a plant is inactive or has no data yet, its days are returned as zero rather than as an error.

How Forecasts Are Built

The forecast rests on two inputs and a physical model that ties them together.

Weather Data

Forecasts use each site's local weather forecast via Open-Meteo, which automatically selects the best regional high-resolution weather model for the location — for German sites typically the ICON-D2 model of the German Weather Service (Deutscher Wetterdienst, DWD). The underlying weather data arrives at 15-minute resolution and is shown on an hourly grid across the 5-day window. Because the forecast is geo-specific to each plant's location, neighboring plants can show meaningfully different outlooks. See Connected Microservices for how external weather data reaches the platform.

Physical Model

Expected production is derived from first principles — a clear-sky model of the site, the plant's installed capacity, and its recently measured performance — not from a learned statistical model. This is the same physics-based foundation used by the Digital Twin for clear-sky and performance analysis, which is why the forecast's clear-sky reference lines up with the monitoring view.

Wind Plants

Wind plants get the same 5-day outlook as solar plants. Instead of sunlight, the wind forecast follows the wind at each turbine's rotor and reads the turbine's own power curve.

How the Wind Forecast Is Built

• Wind at hub height — the weather forecast delivers wind speeds at several heights above ground (10, 80, 120 and 180 m). For every quarter-hour Mirox fits the wind profile between these heights and reads it at the hub height of each turbine. The hub height comes from the turbine drawer or from the turbine's datasheet; if neither has one, 120 m is assumed.
• Air density — at the same wind speed, cold dense air carries more energy than warm thin air. Mirox corrects the hub-height wind for the air density at the rotor, derived from the forecast temperature at hub height and the site's elevation, relative to the reference density of the turbine's datasheet (1.225 kg/m³ unless the datasheet states another value).
• Power curve — the corrected wind speed is read off the turbine's datasheet power curve: no output below the first point of the curve, none at or above the cut-out wind speed, and never more than the rated power. Any datasheet curve is used, certified or not. A turbine without a power curve is left out, and the plant forecast covers the remaining turbines.
• Plant total — the turbine forecasts are added up. Hidden turbines are excluded; every other turbine is assumed available, so planned maintenance, faults and curtailments are not deducted. Where the plant's grid connection limits feed-in, the total is also available capped at that limit.

Unlike the solar forecast, the wind forecast is not calibrated with a performance ratio — it shows what the turbines can produce according to their power curves. There is no clear-sky reference and no clear-sky quality rating for wind.

Where the Wind Forecast Appears

• Forecast page — wind plants are listed next to your solar plants with the same 5-day strip. Their weather chips include the forecast wind speed at hub height.
• Turbines page — the power graph of a wind plant carries a forecast lane, you can page up to four days ahead, and the 5-day strip sits above it.
• Monitor, dashboard and mobile app — wind forecasts fill the same places as solar forecasts, and fleet forecast totals include the energy expected from your wind plants.

Expected Power from the Actual Weather

The same model also runs on the actual weather once it has settled, about six hours behind real time. This expected power continues seamlessly into the forecast, so history and outlook form one line. On wind charts it is shown next to the possible power reported by the turbine controllers, not instead of it: the controllers derive their possible power from the wind measured at the nacelle, Mirox from the weather at hub height. Two independent references make it easier to tell a weak-wind day from a turbine that underperforms.

Forecast Applications

The 5-day window is built for near-term operational and commercial decisions:

• Operational planning — schedule site visits and maintenance during expected low-production days.
• Storage coordination — plan charging and discharging around the expected production curve.
• Grid and market coordination — anticipate feed-in and align with grid operations.
• Revenue anticipation — translate expected production into expected income alongside your Market & Tariffs configuration.

Related Features

• Reports — accounting-grade production and financial reports for completed periods
• Digital Twin — the physics and clear-sky analysis the forecast builds on
• Market & Tariffs — turn expected production into expected revenue
• KPI Dashboard — where the forecast overview surfaces alongside live performance
• Real-Time Monitoring — the measured data that grounds each forecast